NOU 2012: 16

Cost-Benefit Analysis

To table of content

7 Net wider impacts of transportation projects

7.1 Introduction

From the terms of reference of the Committee:

The Committee shall assess how gains from, for example, transportation investments should be dealt with in cost-benefit analysis, including benefits that are currently often not assigned a price tag in cost-benefit analysis, such as productivity effects from increased geographic density, increased labour supply, as well as the interaction between transportation services and land use. The Committee shall assess how the framework may potentially be made more specific when taking into consideration any net contribution from wider impacts of a public transport measure.

A cost-benefit analysis will normally summarise, in monetary terms, the effects for economic agents that are directly affected by a measure. It is therefore appropriate to limit the analysis to effects in the market where the measure is implemented. However, certain projects may have wider impacts of some importance in other markets. If these wider impacts contribute to net value added, and not only represent a redistribution of such value added, one should examine how such effects are to be dealt with in cost-benefit analysis, both as a specific method and in the broader sense attributed to that term in this Report.

A general definition of net wider impacts is provided in Chapter 7.2. Later in the Chapter, however, the Committee will focus, in line with the terms of reference, on wider impacts in the transportation sector. Wider impacts of transportation projects are the subject of ongoing economic research, and the Committee will in this Chapter review current research and discuss the implications for the cost-benefit analysis of transportation projects. The Committee also refers to the general discussion of wider impacts in the NOU 1997: 27 Green Paper.

The Chapter starts out by clarifying some concepts (7.2). This is followed by a presentation of the theoretical (7.3) and empirical (7.4) basis for the analysis of net wider impacts. In Chapter 7.5, we review recommendations from other countries with regard to the treatment of wider impacts in the analysis of transportation projects. Chapter 7.6 provides a brief discussion of other sources of deviations between actual benefits and economic effects as typically estimated. Finally, we present the assessment (7.7) and recommendations (7.8) of the Committee.

7.2 Some concepts

Generally speaking, wider impacts of public projects may be defined as effects in other markets than those directly affected by the measure under analysis. Wider impacts may be both positive and negative. If we use the term primary markets for markets in which the project has direct effects and the term secondary markets for markets in which the project has indirect effects, we may define wider impacts as changes in resource use caused by changes in secondary market equilibria. As far as transportation infrastructure investment is concerned, the transportation markets are the primary markets, whilst the labour market, the property market and the markets for those goods and services that make use of, or are otherwise affected by, transportation services are examples of secondary markets.

Net wider impacts

The Committee will focus the discussion on wider impacts that are of net economic value to the country, and we define these as “net wider impacts”.1 These will involve, in line with the definition above, situations in which the changes in the secondary market equilibria have an impact on economic efficiency. Assuming an economy without market failure, it follows from the welfare theorems that all economic effects of a marginal project will be captured by a well-specified ordinary cost-benefit analysis in the primary markets.2 In order for a wider impact to be of net economic value there must, in other words, be a market failure in the secondary markets that implies under- or overconsumption of resources, relative to the economically optimal resource allocation, in the situation prior to the implementation of the measure.3 If such under- or overconsumption is affected by the measure under analysis, there is a net wider impact from the said measure that may have economic efficiency implications.4 This is often termed “wider economic impacts”. In the present Report we will stick to the term “net wider impacts”. See Chapter 5 of the NOU 1997: 27 Green Paper for a more detailed discussion of various reasons for market failures, as briefly summarised in Box 7.1.

Textbox 7.1 Market failure

Different reasons for market failure are discussed in Chapter 5 of the NOU 1997: 27 Green Paper. The green paper discusses a main finding in economic welfare theory, called the first theorem of welfare economics, to the effect that an economy with free competition in all markets will, under certain assumptions, provide a resource allocation that results in efficient resource use. If the assumptions are not met in any given market we may, generally speaking, conclude that there is market failure in such market. The NOU 1997: 27 Green Paper notes the following causes of market failure:

  • Public goods

  • Externalities

  • Imperfect competition

  • Taxation

  • Disequilibrium

  • Asymmetric access to information.

Wider impact or simply redistribution?

One will in many contexts be interested in the wider impacts of a project locally. This may for example relate to the impact of a transportation project on a specific area. If these wider impacts in one area are mirrored by corresponding, but reverse, wider impacts elsewhere, such wider impacts represent a pure redistribution effect, and not a source of added economic profitability in a cost-benefit analysis context. If goods and services affected by the measure are priced correctly, and there is no market failure in the secondary markets prior to implementation of the measure, effects in other markets will only represent the redistribution of the original effect of the measure, and a well-specified ordinary cost-benefit analysis will capture all relevant economic effects in the primary market. In other words, there will be no net wider impact. If the redistribution effects are considerable, they may nonetheless be of relevance to the assessment of the decision maker, and should therefore be discussed in the analysis if possible; see Chapter 3 on distribution.

7.3 Net wider impacts: Theoretical background

The terms of reference identify three categories of benefits that are currently often not assigned a price tag in the cost-benefit analysis of transportation projects: productivity effects from increased geographic density, increased labour supply in the presence of distorting taxes or involuntary unemployment, as well as the interaction between transportation services and land use.

As discussed in Chapter 7.2, one is only faced with a net wider impact if the measure under analysis has an impact on any under- or overconsumption resulting from a market failure. It is therefore necessary to look at what types of market failure may give rise to these potential additional efficiency effects. In the present Chapter, we will examine how transportation projects in an area with unexploited economies of scale may give rise to productivity effects from increased geographic density (7.3.2), as well as how a transportation project in an economy with distorting taxes may give rise to an efficiency effect from increased labour supply that is not captured by the market agents (7.3.3). Thereafter, we will discuss the extent to which the interaction between land use and transportation measures may entail a net wider impact (7.3.4), as well as the effects of transportation projects in markets with imperfect competition (7.3.5). Finally, we will look at some estimation methods (7.3.6). However, first we will briefly outline, to put the discussion in context, what will already have been captured by a well-specified ordinary cost-benefit analysis in the transportation market. The main purpose of this is to highlight the fact that some effects that may be perceived as wider impacts will in actual fact already have been captured by an ordinary analysis of the transportation market.

7.3.1 What is captured by an ordinary cost-benefit analysis of a transportation project?

It may be useful, before the Committee addresses sources of net wider impacts in detail, to outline more specifically how analyses of the effects of a transportation project are currently conducted. In order to make the presentation as specific as possible, we will look at the Norwegian Public Roads Administration’s cost-benefit analysis guidelines (Norwegian Public Roads Administration, 2006). These guidelines describe an approach under which all costs and benefits are distributed across four main stakeholder groups:

  • Road users and transport users

  • Operators (public transport providers, parking companies, road toll collectors and other private sector stakeholders)

  • The public sector

  • The rest of society (accidents, noise and air pollution, residual value, tax funding cost).

The economic effect of the measure is the sum of the effects for these four main groups. The effects are valued by using calculation prices. The effects may arise at different points in time, and are measured at net present value in order to make the values comparable. Any transfers between the groups will be zeroed out in the final sum, and one ends up with an expression for the net economic implications of the priced effects of the measure.5 The non-priced effects are discussed separately.

In order to clarify this approach, let us look at the example of a road project that reduces the travel time between two locations. We assume, for the sake of simplicity, full public funding and no public transport. The costs faced by road users and other transport users when considering whether to travel are termed “generalised travel costs” and include time costs, fuel expenses, tolls, etc. (in addition to costs like bus tickets, ferry tickets, etc., when public transport is included). A transportation project reducing the travel time between two locations will produce an economic benefit effect through reduced generalised travel costs. In addition, lower costs may result in somewhat higher traffic, inasmuch as road users who previously thought, for miscellaneous reasons, that the journey was not worth the cost will now want to take the journey. The overall effect of the transportation measure will depend on the magnitude of the change in travel costs, the number of road users prior to the implementation of the measure and how the road users change their behaviour as the result of the change in costs. If the appropriate calculation prices are used, this will represent the changes in the consumer surplus of users. These benefits on the part of users must be balanced against the direct costs associated with the measure and the valuation of the effects for the rest of society. This latter category includes, for example, the value of the change in the number of expected accidents, the value of emission changes and the economic cost of taxation. Since different effects occur at different points in time, the values are converted into a net present value.

The method described above analyses the effects of a measure through its direct effects in the transportation market and some specific effects for the rest of society; what are termed the primary markets in Chapter 7.2. The effects may over time also be reflected in other markets, i.e. the secondary markets. However, unless one is facing a market failure in the secondary markets that may give rise to net wider impacts, this is only a reallocation of resources – i.e. a redistribution effect, as explained in Chapter 7.2. If one includes both the direct benefits in the transportation market and the effect of such redistribution of benefits between the primary and the secondary markets, one is going to count the same thing twice. This is discussed is more detail in Box 7.2.

Textbox 7.2 Redistribution of benefits under perfect competition

Chart 7.1 provides a schematic representation of the economic benefits from a road project. This pertains to a project that reduces the generalised travel cost for individuals by shortening the distance travelled, thus delivering travel time savings. We analyse the issue in a partial equilibrium model under the assumption of perfect competition.

Figure 7.1 The effect of a road project for road users and transport users.

Figure 7.1 The effect of a road project for road users and transport users.

The demand curve indicates the private marginal willingness to pay for the relevant journey. The horizontal axis shows the demand for travel and the vertical axis shows the total cost per journey faced by the road user, including expenses associated with the means of transport and the value of time lost (generalised travel costs).

Source Norwegian Public Roads Administrations (2006), Handbook 140.

The cost of taking the journey is measured along the vertical axis. The benefit to an individual from taking the journey is reflected by the demand curve, which shows the private marginal willingness to pay. An individual will take a journey if her benefit exceeds the cost she is facing. At travel cost G0, X0 journeys will be completed. The project we are analysing will reduce the cost of taking the journey from G0 to G1. Those who were already taking the journey prior to the implementation of the project will see a reduction of their travel costs of exactly G0- G1. This will result in a higher user surplus on their part. Moreover, some people who did not travel prior to the project will now be taking the journey, since the cost has declined. The number of new journeys is represented by the change from X0 to X1. The result will be an overall increase in consumer surplus, represented by the area between the marginal willingness to pay curve and the new travel cost. For the first new traveller, the benefit will be slightly less than (G0-G1), whilst the benefit will be just above zero for the last new traveller. The point (G1, X1) is the new equilibrium subsequent to the implementation of the project.

We note from such a partial analysis that the change in generalised travel costs reflects an upper limit with regard to individual willingness to pay for a transportation project. This implies that, for example, the benefit of improved proximity to one’s customers upon the construction of a new bridge will be represented by the decline in generalised travel costs. We note, by reference to Chart 7.1, that it is necessarily the case that a firm’s benefit cannot exceed the change in such firm’s generalised travel costs resulting from the project. If the benefit had been higher, the trade would have taken place also before the bridge was constructed. The resulting benefit on the part of both the firm and the customers is the reduced generalised travel cost included in an ordinary cost-benefit analysis. If one had included a value attached to improved proximity to customers, in addition to the benefits to road users resulting from reduced generalised travel costs, one would have double counted the benefits.

The reasoning outlined immediately above is equally valid when the effects of a project are reflected in changes to other prices, for example land prices. Assume that the benefits from the above bridge project are mirrored by an exactly matching increase in housing prices on an island that has been linked to the mainland as a result of the new bridge. The overall benefits available for distribution remain unchanged, but will now accrue to island property owners rather than to road users. For cost-benefit analysis purposes, it will in this case be appropriate to calculate the project benefits on the basis of the benefits to road users from reduced generalised travel costs, irrespective of how such benefits are distributed. However, if the distribution effects are deemed to be significant, these may be outlined separately, cf. the discussion of distribution effects in Chapter 3. The interaction between land use and transportation is discussed in Chapter 7.3.4.

7.3.2 Productivity and geographic concentration of economic activity

Sources of increased productivity

Economic geography has provided new and specific insights into which mechanisms influence the interaction between economic activity and the location in which such activity is taking place. It is hardly a new observation that location is of importance to economic behaviour and efficiency. As early as in 1890, Alfred Marshall wrote about the economies of scale existing in towns and cities in his book Principles of Economics, as exemplified by the following quote:

When an industry has thus chosen a locality for itself, it is likely to stay there long: so great are the advantages which people following the same skilled trade get from near neighbourhood to one another. The mysteries of the trade become no mysteries; but are as it were in the air, and children learn many of them unconsciously.

In addition to such dissemination of knowledge in networks, Marshall noted the advantages of a well-developed labour market and the beneficial effect of good interaction between manufacturers of finished products and their suppliers.

The effects noted by Marshall in 1890 are also the focus of current research into how productivity and employment are affected by geographic density. Such effects of the geographic concentration of economic activity may be termed agglomeration effects or proximity effects. Different theoretical models seek to explain how proximity between agents, most often expressed by a measure of functional town size, may influence productivity. A joint feature of these explanations is that they are predicated on various forms of market failure, which were also highlighted in the NOU 1997: 27 Green Paper as prerequisites for taking into consideration the resource reallocation resulting from a public measure. Special mention is made of various forms of economies of scale that are not fully utilised in towns.

Duranton and Puga (2004) provide an overview of the literature addressing the microeconomic mechanisms that may give rise to such economies of scale. The economic mechanisms that may create a causality between functional town size and productivity include:

  • More well-functioning labour markets. Contribute to a better match between the knowledge and skills needed by employers and the knowledge and skills offered by employees.

  • Linkages to both upstream and downstream markets. By linkages to upstream markets are meant better-functioning factor markets, whilst linkages to downstream markets are to do with more efficient markets for finished products.

  • Dissemination of knowledge in networks. Co-location of enterprises in so-called clusters may contribute to enhanced productivity within the region.

Duranton and Puga (2004) believe the literature within the field to be sufficiently mature to allow a few general conclusions to be drawn. They note, inter alia, that a positive correlation between functional town size and productivity is consistent with a wide range of microeconomic models. This suggests that such a correlation is theoretically robust. It is, at the same time, difficult to empirically identify which specific mechanism produces the observed outcome.

Implications for cost-benefit analysis

Venables (2007) discusses cost-benefit analysis in a situation where a correlation between functional town size and productivity is assumed to exist. He notes that if an infrastructure investment in a town results in more employees being able to commute into town, the implication is a functional enlargement of that town. This will, according to the correlation presumed in theory, result in enhanced productivity, not only for the new arrivals, but for all employees in town. The value of the productivity enhancement will be additional to the sum total of changes in travel costs on the part of those travelling on the affected route. This represents a positive net wider impact of the transportation measure. The interaction with distorting taxes is also noted. This is discussed in more detail in Chapter 7.3.3.

There are three limitations to the above arguments, according to Venables (2007). Firstly, the reasoning applies to commuter journeys only, and such effects will, in a cost-benefit analysis that also includes leisure journeys, need to be weighted by the portion accounted for by commuter journeys. Secondly, the analysis has failed to examine the relationship between transportation improvements and negative externalities from increased crowding and queuing. Such effects will be the reverse of positive wider impacts. Thirdly, the model assumes that productivity outside the town remains unchanged despite labour being relocated to the town, which assumption is not necessarily correct. Caution is urged when it comes to making direct use of the findings to render policy recommendations, but one of the benefits of infrastructure investments in town may be overlooked by disregarding said findings.

7.3.3 Labour supply increase in the presence of distorting taxes

A tax on income implies that individuals make decisions based on their wages after tax, whilst the value of their production to society is its economic value, which in a well-functioning market will be reflected in wages before tax. This is a market failure with implications for the evaluation of transportation projects.

Reduced travel time may result in expanded leisure hours, expanded working hours or a combination of the two. It is conceivable that more people will enter the labour market when travel costs have declined. If a transportation project results in increased labour supply, it will be accompanied by higher production and more value added.

Part of the value associated with increased labour supply is captured in a well-specified ordinary cost-benefit analysis of the transportation market. Since the transport users who increase their labour supply will make their decisions on the basis of wages after tax resulting from such increased labour supply, this effect will be captured as the change in the consumer surplus of the transport users.

However, the value of the production increase resulting from the expanded labour supply is equal to wages before tax. In order for the entire value of expanded labour supply to be included in the analysis, one needs to add such part of the value of the production increase as is not captured by an ordinary cost-benefit analysis of the transportation market. Such additional element corresponds to the difference between the economic value of increasing production by one additional hour of work and the value received by individuals if they work for one additional hour. In practice, such difference is equal to the additional tax revenues resulting from the change in labour supply. This constitutes a net wider impact. See Box 7.2 for a graphic representation of this effect.

Textbox 7.3 Economic value of increased labour supply

The tax wedge implies that if reduced travel costs result in increased labour supply, there will be an additional effect that is not captured by an ordinary analysis of the transportation market. This is illustrated in Chart 7.2.

Figure 7.2 Economic effect of increased labour supply in an economy with a tax on income.

Figure 7.2 Economic effect of increased labour supply in an economy with a tax on income.

Chart 7.2 presents the difference between the economic effect of increased employment and such part thereof as will be captured by an ordinary cost-benefit analysis. The number of people in employment (L) is represented by the horizontal axis, and the wage (w) by the vertical axis. We assume that the supply of labour (T) is an increasing function of wage. This function represents the value of the leisure that must be forgone in order to participate in working life. In order to keep the illustration simple, we disregard the fact that wages can be expected to be a declining function of the number of people in employment, and assume that the wage curves are horizontal.

In deciding whether to participate in working life, an individual makes a trade-off between what she earns by working and the value of the leisure time she must forego when working. If the wage after tax, net of commuting costs, exceeds the value of leisure time, she will work. If the wage after tax, net of transportation costs, is less than the value of leisure time, she will not work. To begin with, labour supply will be L0, given the wage after tax, net of transportation costs. If we implement a project that eliminates the transportation cost, the labour income of individuals will increase by the amount of the eliminated transportation cost. The new labour supply is L1.

What is the economic effect of the transportation measure? The disposable income of those who were already participating in the labour market will be higher after the elimination of the transportation costs. This is represented by area A in Chart 7.2. Moreover, the increase in labour supply, either by those who participated in working life increasing the number of hours worked or by new people entering the labour market, will benefit the employees involved, inasmuch as what they earn from work exceeds their value of leisure. This is represented by areas A and B in Chart 7.2. These effects will be captured by an ordinary cost-benefit analysis. However, the value of the production generated by the increase in labour supply corresponds to the wage before tax. The difference between the economic value of increased employment and the value reflected in individual decisions is represented by area C in Chart 7.2. This corresponds to the value of the increase in tax revenues resulting from increased employment, and represents a net wider impact of increased employment in an economy with a tax on labour income.

It is an empirical question whether transportation projects result in a net increase in labour supply, by more people entering the labour market or by people working longer hours, and we will revert to this in Chapter 7.4.

Venables (2007) also looks at how this labour market effect interacts with the productivity increase resulting from increased functional town size, cf. Chapter 7.3.2. In his model, the combination of lower generalised travel costs and enhanced productivity in town will result in more people choosing to work in town rather than in the surrounding rural areas. Individuals will choose to commute into town as long as the wage difference exceeds generalised travel costs. However, employees will make their decisions based on wages after tax, and not on the economic value of production, which is measured by wages before tax. Hence, the additional tax income resulting from employees choosing to commute into town to get more productive jobs reflects a production increase that would not have occurred otherwise. This pertains to the change in tax revenues from those employees who choose to commute into town as a result of the measure. The model in Venables (2007) assumes that labour supply in the country is fixed. However, higher wages in town may give rise to a countrywide increase in labour supply, and not only to a reallocation of labour from less productive jobs outside town. An analysis of overall increases in labour supply must, because of distorting taxes, take into account the fact that individuals make their decisions on the basis of the wage after tax, whilst the value of their production is its economic value. In a well-functioning market, the latter is reflected in the wage before tax.

7.3.4 Land use and transportation

A transportation project may result in price changes in the property market. This will, generally speaking, only represent a redistribution of the original direct benefits from the measure, as captured by the benefit to road users in a well-specified ordinary cost-benefit analysis. Hence, including both effects in the analysis would amount to double counting.

If a measure releases an area that was previously used for transportation, and such area has a positive scarcity value, its value in its best alternative use should in principle be included on the benefit side of the project. This is discussed by Minken (2011). It is there argued that this represents a real economic effect that is not captured in the direct user benefits from the project. It is emphasised that in order for such effect to merit inclusion in the analysis, the probability of such area being entered into use must be high, e.g. through serious expressions of interest. Minken (2011) notes that methods for such valuation are available in Sweden6, although these should not be used indiscriminately.

Another key aspect of the interaction between land use and transportation measures is how increased geographic density may entail enhanced productivity, cf. the discussion in Chapter 7.3.2. If one envisages that a transportation project will increase productivity in an urban area through the mechanisms discussed in Chapter 7.3.2, one may also be faced with a situation in which the value of enhanced productivity triggers higher property prices in the urban area. However, if one has sought to estimate the value of enhanced productivity directly, it would amount to double counting if one were to include the effect of the property market price changes in addition thereto.

If the de facto rationale behind a transportation measure under analysis is, for example, that it offers new urban development opportunities, such measure should be evaluated from that perspective in a cost-benefit analysis. It would be appropriate for the said analysis to examine which land use opportunities the measure will facilitate in both the short and the long run, and the potential economic effects resulting therefrom. In such case, economic effects in the transportation market will be only one aspect of the analysis. A recent example of an urban development project is the rearrangement of the traffic system at Bjørvika in Oslo, which project involved, inter alia, redirecting traffic underground. The net economic benefit estimate was here restricted to the actual road project only, which benefit was estimated at NOK -2.9 billion in the proposition submitted to the Storting in 2005, based on the priced factors. The expected cost of the project was about NOK 3.9 billion at 2005 prices. It was emphasised in the funding proposition to the Storting (Proposition No. 50 (2004-2005) to the Storting) that the benefits from the project’s facilitation of urban development were not included in the net economic benefit estimates. One may say that the Storting, by giving the go-ahead for the Bjørvika road construction project, valued the non-priced factors as more than compensating for the negative net benefits estimated from the priced factors.

7.3.5 Imperfect competition

Imperfect competition in the markets that make use of transportation services may also be a source of net wider impacts. Firstly, improved transportation may increase competition between enterprises, which may narrow the gap between prices and opportunity costs. High transportation costs can be a main cause of geographic division in some markets, thus resulting in a large number of small, local markets. To the extent that a measure does actually have an impact on competition, the resulting economic effect will not be captured by ordinary cost-benefit analysis.

Secondly, there is scope for net wider impacts of increased production in markets with imperfect competition even in the absence of competition improvements. This is discussed in, inter alia, a report from the UK Department for Transport; SACTRA (1999). Its approach parallels the one discussed in Chapter 7.3.3, and focuses on effects of the under-consumption of resources as the result of market failure or distorting taxes.

Figure 7.3 Efficiency benefits in markets with imperfect competition.

Figure 7.3 Efficiency benefits in markets with imperfect competition.

Let us look at a firm operating in a market characterised by imperfect competition, which firm may therefore charge a price in excess of marginal cost. Assume also that the said firm is facing transportation costs that vary in proportion with its production level, thus implying constant marginal costs. In Chart 7.3, this is illustrated by a monopolist, and the profit-maximising price and output of the monopolist is given by the point Pm, Qm. If a transportation project results in lower transportation costs for the firm, this will, all else being equal, mean lower marginal costs for the firm, illustrated by a shift in the marginal cost curve from MC1 to MC2. The firm responds to the lower production costs by reducing its price and increasing its production to P2, Q2. The area A+B corresponds to the benefits captured as the value of reduced generalised travel costs for the business sector in an ordinary cost-benefit analysis. We note from Chart 4.3 that, since the firm is able to charge a price in excess of marginal costs, consumers’ willingness to pay for one additional unit (as illustrated by the demand curve) exceeds the cost of producing such unit (the marginal cost curve). When production is expanded, this results in an efficiency benefit that is additional to that captured by an ordinary cost-benefit analysis. Such additional efficiency benefit is illustrated by the area C+D+E in Chart 7.3. This efficiency benefit is not captured by the firm and represents net wider impacts of expanded production in markets with imperfect competition.

7.3.6 Estimation methods

The transportation sector in Norway has commissioned two reports that discuss the sources of net wider impacts (Heldal et al., 2009, COWI, 2012). The reports also outline potential methods for estimating net wider impacts, but emphasise the need for further improvement of estimation models, as well as the need for empirical studies and basic data for purposes of making actual calculations operational. The reports present some case studies highlighting the abovementioned need.

In principle, a study of a transportation project in a general equilibrium model will be an alternative to a partial cost-benefit analysis in which one seeks to take net wider impacts in other markets into account. If the general equilibrium model does comprise various forms of market failure, it may be able to capture any net wider impacts. Consequently, spillover effects between sectors through market interactions may result in government intervention and market failure in other parts of the economy than those where the measures are implemented having an impact on their economic costs and benefits (Fæhn, 2010).7 Work has been conducted to expand general equilibrium models by adding realistic transportation networks. These are generally termed Spatially Computable General Equilibrium (SCGE) models. Hansen (2011) refers to a number of operational models of this type in other countries than Norway that can be used to estimate net wider impacts of transportation projects.8 However, the level of detail required to model a complex transportation market within such SCGE models may make it difficult to identify the various effects in each market. On the other hand, not including sufficient details about the transportation system in the model may result in the model being unable to capture the relatively minor changes to the transportation system represented by many individual projects (Minken, 2011).

7.4 Empirical analyses

The NOU 1997: 27 Green Paper concluded, inter alia, that strict requirements must be met by the empirical basis in order for a cost-benefit analysis to capture net wider impacts. It was noted that improved underlying data, enhancement of econometric methods, etc., may over time enable the documentation of wider impacts in areas where such effects were difficult to identify at the time of the green paper, and that such might be the case within the transportation sector.

A considerable number of empirical studies of wider impacts have been conducted since 1997. Expanded data availability, enhancement of econometric methods and, not least, the possibility of processing large data sets quickly and cheaply by using more powerful computers, have paved the way for several interesting studies. Most of the relevant studies have been carried out outside Norway, and are reviewed in Chapter 7.4.1. Studies based on Norwegian data are reviewed in Chapter 7.4.2.

7.4.1 Studies of net wider impacts of transportation projects outside Norway

Productivity and economies of scale

The effects of larger functional town size on productivity within an area have been addressed by a number of studies. These studies typically seek to estimate by how many percent productivity in a town increases if actual town size is increased by one percent (the elasticity of productivity with respect to a measure of functional town size).

Melo et al. (2009) conclude, in a comprehensive review of available empirical research within the area, that the findings exhibit major variations.9 The study shows that estimates vary widely depending on sector, country and how functional town size is measured. Besides, there are variations in the different methodological choices made in these studies. The mean elasticity reported by Melo et al. (2009) is 0.058. This suggests that if functional town size increases by one percent, productivity in such town will increase by 0.058 percent. Most of the elasticities fall within a range from -0.090 to 0.292.10 The lowest elasticity identified by Melo et al. (2009) is -0.8, i.e. a negative correlation between town size and productivity, and the highest is 0.658, which shows that the estimates are widely dispersed. The mean elasticities reported in studies also vary between countries. The largest number of estimates is from the United Kingdom, where the mean elasticity is 0.102. A mean of -0.038 is reported for Europe as a region, whilst 0.039 is reported for France and 0.017 for Sweden. There are also considerable variations between sectors. The authors find a mean elasticity of 0.040 for the manufacturing sector, whilst the mean elasticity for services is 0.148. The article does not include any studies using Norwegian data.

Melo et al. (2009) conclude that the correlation between productivity and functional town size should be evaluated in the specific context in which the estimate is to be used. They take the view that there is no reason to expect elasticities to be the same across sectors, towns or countries.

The findings of Melo et al. (2009) are in line with a number of other studies within this field. Vickerman (2010) notes, for example, that such empirical studies will typically find elasticities ranging from 0.01 to 0.1, and highlight variations between sectors and regions. Furthermore, a more recent study of OECD countries (Égert et al., 2009) finds major variations in both the direction and the magnitude of the correlations identified between transportation investments and productivity, concluding that the effects are country-specific and depend on, inter alia, the level of infrastructure development in a country.11 Graham et al. (2009) also find major variations between sectors in a study based on UK data, with an elasticity for manufacturing of 0.021, construction of 0.034, consumer services of 0.024 and producer services of 0.083.

There are also studies showing that the effects of proximity decline quite rapidly with distance. The question is how close different locations with economic activity need to be in order for economic activity in one location to influence such activity in the other location. Graham et al. (2009) has examined how rapidly the effect declines. Based on UK data, they find, for example, that the effect of the employment level in one area on productivity in another area declines by more than the inverse of the distance.12 This implies that the effect from employment 10 kilometres away from the town centre is less than 1/10th of the effect from employment one kilometre away from the town centre. They have also distinguished between effects from economic activity on the basis of selected distance intervals, and find that correlations are positive and statistically significant between 0 and 25 kilometres, and then declining in the 25-to-50 kilometre interval. No statistically significant correlation between proximity and productivity is identified for distances in excess of 50 kilometres.

A review of the literature relating to agglomeration and public transport (Chatman and Noland, 2011) concludes that it is uncertain whether, and in which cases, public transport improvement may entail enhanced productivity as the result of efficient urban concentration. The review notes that very few studies have specifically addressed the effect of public transport.

There are methodological problems associated with identifying the effect of increased functional town size on productivity. It is noted that it is difficult to know whether enhanced productivity is caused by improved accessibility within an area, or whether it is the case that authorities and other stakeholders improve accessibility within an area characterised by productivity enhancement (see e.g. Vickerman, 2010). Furthermore, it is noted that the way in which a study is designed has a major impact on the elasticity estimates (Melo et al., 2009). Chatman and Noland (2011) note that it is not sufficient to identify a correlation between agglomeration and productivity, as many studies have done. What is relevant is the correlation between transport and agglomeration and, thereafter, how transportation-generated agglomeration influences productivity. It is noted that no empirical studies have examined such chain of causality. Chatman and Noland (2011) also note the need for examining any agglomeration effects in the context of the effects from any reduction in the density of economic activity elsewhere, since such a net effect is the relevant parameter at the national level. Graham and van Dender (2010) emphasise that even large projects have little impact on the overall functional town size of a given town, and argue that traditional approaches to estimating wider impacts are not paying sufficient heed to this. In addition to the methodological problems mentioned above, the correlation between productivity and functional town size is not necessarily linear. Graham and van Dender (2010) show that elasticity estimates vary with how geographically concentrated economic activity is to begin with, and that no correlation can be identified for certain segments. They therefore conclude that it is not possible to distinguish the effect of functional town size on productivity from other potential explanations13. It is noted that one direct implication for the cost-benefit analysis of transportation projects is that it would be highly misleading to use simple point estimates to represent wider impacts in such an analysis.

Employment

A number of studies carried out at macro level have also examined the correlation between road capital and employment. Vickerman (2008) notes that no systematic positive correlation has been found between employment and infrastructure, especially when distinguishing between different road types. It is also noted that what might at first glance seem to be a positive correlation may reflect methodological errors in the statistical analysis. One such example is Jiwattanakulpaisarn et al. (2009), which analyses the effect of varying freeway network densities in different counties in the US state of North Carolina. By looking at the relationship between freeway network densities and employment, they found a positive correlation, but when taking into account that the causality might be the reverse, and that employment in one period might be determined by employment in the previous period, they no longer found any such correlation.14

Imperfect competition

The UK Department for Transport (2005) discusses potential effects of transportation projects under imperfect competition. It is noted that there is limited empirical material analysing the correlation between transportation investment and the degree of competition in the economy. One approach taken is to examine the competition implications of changes in barriers to trade; it is estimated, according to the Department for Transport (2005), that a five percentage-point reduction in the average rate of customs duties will result in a 4.5 percentage-point reduction in mark-ups in the economy as a whole. Assuming that transportation costs represent five to seven percent of the overall costs of firms, this finding suggests that a 70 to 100 percent reduction in transportation costs is needed in order for a transportation measure to have a corresponding impact on competition. In practice, transportation measures will have much less of an effect on transportation costs than do barriers to trade. It is concluded, on this basis, that one cannot normally expect a transportation project to have any net wider impacts as the result of enhanced competition.

Attempts have been made at quantifying net wider impacts resulting from increased production in markets characterised by imperfect competition. Under certain assumptions, the net wider impact resulting from increased production under imperfect competition may be captured by applying an “uprate factor” to the value of reduced generalised travel costs to the business sector in an ordinary cost-benefit analysis (Department for Transport, 2005). The magnitude of this “uprate factor” will depend on the size of the gap between price and marginal cost, as well as the elasticity of demand in the affected secondary market. If the gap between price and marginal cost is large and demand is elastic, this effect may be of importance. If the gap between price and marginal cost is small and demand is inelastic, the effect may be less important. It is emphasised that such a method of calculation can only be a rough approximation.15 Based on studies of the average gap between price and marginal cost, as well as the mean elasticity of demand in the UK economy, the Department for Transport (2005) concludes that one will capture net wider impacts resulting from increased production under imperfect competition in the United Kingdom by adding 10 percent to the commercial transportation benefits in a cost-benefit analysis. It is noted that the simplification is only valid if the net wider impacts resulting from imperfect competition are the same for all sectors and the transportation project has the same impact on all sectors.

Other studies

The issue of whether infrastructure development stimulates increased value added has been analysed earlier. Aschauer (1989) compared production and productivity with infrastructure investment levels over time and in different regions of the United States. He noted a positive correlation. The NOU 1997: 27 Green Paper identified methodological weaknesses in the approach, as also emphasised by Vickerman (2010). He notes that doubt as to the direction of the causalities, in particular, means that the method used by Aschauer (1989) now inspires little confidence, and that the said approach is especially problematic as a basis for cost-benefit analysis. The studies also disregard the fact that economic agents may adapt to different transportation cost levels by changing their investments in homes and activities, or through expanding into new markets for finished products and inputs.

Some studies also look at behavioural changes at the micro level caused by transportation projects. As an example, Gibbons and Machin (2005) identify, according to Vickerman (2010), a systematic correlation between housing price increases and station access improvements on a new underground line in London (Jubilee Line). Such geographic rearrangement of activity as the result of an infrastructure investment is not a net wider impact in itself, but may give rise to the type of cluster effects and labour market effects discussed in Chapter 7.3.2.

A number of studies have also been conducted on the effect of the French high-speed trains. In summary, Vickerman (2010) notes, in a review of such studies, that the establishment of high-speed trains has resulted in increased traffic between the towns involved, whilst the overall economic effect is much more unclear. It is noted, as a general observation, that the establishment of the high-speed train links between Paris and surrounding towns cannot be said to have resulted in any net redistribution of economic activity between Paris and towns in its vicinity, or to have influenced overall economic growth in these towns. However, it is noted that the establishment of high-speed trains seems to have contributed to centralising economic activity around the cities located along the train routes.

7.4.2 Studies on Norwegian data

Productivity, economies of scale and employment

Studies based on Norwegian data identify no general correlation between road investment and productivity. A macro study based on national accounts data for the period 1963 to 1997 did not, for example, generate findings that supported a hypothesis to the effect that infrastructure investments have improved productivity in Norway.16 Nor does a new study, based on data specified by county from 1997-2005, find any statistically significant correlation between road investments and productivity (Eriksen and Jean-Hansen, 2008). The authors are of the view that the outcome may be caused by statistical problems, but also by the fact that transportation projects in Norway have only to a limited extent been given priority on the basis of their economic profitability.17 Nor did Egert et al. (2009), which analysed the correlation between productivity and infrastructure in OECD countries, find any statistically significant correlation between road/railway investment and productivity in Norway.

Another study of 102 major road projects in Norway that were completed during the 1993-2005 period, finds a positive and statistically significant correlation between investment and population growth, but finds no statistically significant impact on employment, income level, commuting or industrial areas (Lian and Rønnevik, 2010).

The importance of road links between islands and the mainland in stimulating the local business sector is often highlighted in Norway. As a main rule, these effects will represent a redistribution of the regular first-order effects resulting from reduced generalised travel costs, cf. the discussion at the beginning of this Chapter. The Committee has not identified any studies uncovering any wider impacts beyond this, i.e. any net wider impacts. A qualitative study of four firms located in the vicinity of two fjord crossings (links between Giske and Ålesund and between Bergen and Askøy) was unable to identify any additional effect for the business sector in these areas, as the result of improved proximity between business networks (Bråthen, 2000).

The Committee is not aware of any detailed econometric studies of the correlation between town size and productivity on Norwegian data. In a report from 2011, Heum et al.18 note, inter alia, that wage levels in some labour market regions, where regions are defined on a discretionary basis, are higher than in the rest of the country. However, the study is not an empirical study of any causality between geographic economic density and productivity in Norway. Consequently, the findings must be understood as providing an indication as to the hypothetical implications of such a correlation, provided that the assumptions in the said report are in fact met.

Imperfect competition

Klette (1994) found, in a study of imperfect competition in Norway, that the average gap between price and marginal cost within the manufacturing sector in Norway over the period 1980-1990 was five to ten percent. The study found relatively large variations between enterprises within the same industry. This finding indicates that one can only expect fairly minor welfare effects from policy aimed at enhancing competition. Furthermore, he found that no industries were encountering economies of scale.

7.4.3 Empirical evidence - summary

Many studies have been conducted on the correlation between functional town size and productivity. These are of relevance to transportation analysis, since urban transportation projects can be said to increase functional town size by reducing travel time. Most of the studies that have been carried out find a weak positive correlation. However, the consensus within the literature is that there is no uniform correlation across sectors, towns and countries. This implies, inter alia, that the local industrial structure is of major importance in determining which net wider impacts can be expected from a measure. It is also noted that the correlation between productivity and functional town size is not necessarily stable within the same town and sector. Consequently, it will not be correct to use one point estimate to represent the correlation between town size and productivity for cost-benefit analysis purposes.

It is also difficult to distinguish the impact of functional town size on productivity from that of other potential explanatory variables. In practice, this means that a correlation attributed to functional town size in the studies may in actual fact have other, unknown causes. This indicates that one should be cautious about making use of estimates to represent such correlation without having clarified these methodological problems.

The vast majority of the studies have been carried out on UK data, and the correlation found appears to be stronger in the United Kingdom than in other countries. The cause of this is unclear. The Committee is not aware of any studies of such correlation on Norwegian data, but findings from Sweden show a weaker correlation than in the United Kingdom.

Studies on Norwegian data have examined the correlation between road investment and productivity, without specifically addressing functional town size. These studies find no general correlation between road investment and productivity.

A number of studies have also focused on the correlation between road capital and employment at the macro level. There is nothing in the literature to support a general systematic positive correlation between employment and infrastructure. Nor do empirical findings merit a conclusion to the effect that transportation projects will increase competition.

7.5 Recommendations in other countries

Only a small number of countries recommend the inclusion of estimates for net wider impacts in their cost-benefit analysis guides. The United Kingdom stands out by having developed a specific template for such calculations, for purposes of supplementing ordinary cost-benefit analysis. In Japan, the Netherlands, Germany and France, estimates for net wider impacts have been prepared in relation to a select group of very large projects.19

7.5.1 The United Kingdom

Her Majesty’s Treasury recommends, in its general cost-benefit analysis guide (HM Treasury, 2003), examining both the direct effects of a measure and the potential wider impacts of such measure on other parts of the economy. It is noted that such wider impacts need to be clearly described and carefully assessed, since both benefits and costs may be involved.20 The UK Department for Transport recommends, in its guidance materials (Department for Transport, 2012a and 2012b), that the calculation of any net wider impacts should be kept separate from the calculation of net project benefits for cost-benefit analysis purposes. However, calculations of net wider impacts may be included as supplementary information in the summary of the cost-benefit analysis of a measure. The UK Department for Transport has prepared valuation guides for such net wider impacts.

The UK Department for Transport guide discusses four elements relating to the valuation of wider impacts: enhanced productivity as the result of increased functional town size, reallocation of labour to more productive jobs, general labour supply effects and production changes in markets with imperfect competition. The UK Department for Transport notes that the resources devoted to such analysis must be commensurate with the scale of the project under analysis. Consequently, it is for the project owner to evaluate whether one should embark on a detailed analysis of net wider impacts. Production changes in markets with imperfect competition21 and labour supply effects22 are highlighted as net wider impacts that will be of relevance to most projects. Productivity effects caused by increased functional town size23 shall be evaluated if the investment also improves the accessibility of an area included in a list of “functional urban areas” in the United Kingdom. The economic effects of the reallocation of labour to and from more or less productive jobs shall, according to the guide, only be examined if such reallocation is probable on the basis of a detailed, specific model, and shall in any event take the form of a supplementary analysis only, and not be included in the summary of net wider impacts.

In practice, specific attempts at estimating economic net wider impacts have only been made for very large transportation projects. One example is a railway line through London, the Crossrail project, with an estimated cost of about GBP 16 billion, or about NOK 150 billion. The UK Department for Transport estimated that net benefits from the project increased from GBP 12.8 billion to GBP 20.0 billion when including these effects.24 Consequently, estimates for net wider impacts contribute to the project being considered economically profitable. However, Crafts (2009) notes that the Crossrail project in London must definitely be characterised as a project with an unusually strong correlation between infrastructure investment and productivity.

7.5.2 HEATCO (EU)

In its report from 2006, HEATCO, an EU project seeking to harmonise the valuation of transportation projects, discusses elements that may be labelled as wider impacts (HEATCO, 2006). These are referred to as indirect economic effects. The report emphasises that it is important to distinguish between the direct and the indirect effects of a measure in order to avoid double counting. Reference is made to a review of how this is dealt with in 26 countries.25 The said review includes effects referred to as net wider impacts in the present Chapter, but also outright redistribution effects and other aspects of the analysis of transportation market effects. The issues discussed in one or more countries include land use, economic development, short- and long-term employment, interregional equalisation at both the national and the EU level, urbanisation, network effects, effects on government finances and social equality. Moreover, HEATCO notes that the gap between theory and practice within this area is wide, and that its recommendations reflect this. It is recommended that the cost-benefit analysis of transportation projects shall include, at a minimum, a qualitative assessment of potential indirect effects, and that light should be shed on the magnitude of any net contribution to economic profitability on the basis of studies of projects in a similar context. Furthermore, it is recommended to use an economic model to estimate the indirect effects where these are deemed to be of major importance. General equilibrium models featuring a realistic transportation network, so-called Spatially Computable General Equilibrium (SCGE) models, are considered the best alternatives for such an analysis.

7.5.3 Sweden

In Sweden, a committee has recently discussed cost-benefit analysis methods within the transportation sector (Swedish Transport Administration, 2012). A separate chapter discusses how to deal with any net wider impacts. The highlighted elements are largely concurrent with those presented in the present Chapter. It is recommended that no attempt be made at capturing any net wider impacts of small or medium-sized projects. For large projects it is noted that one may present, if needed, a supplementary analysis of any effects beyond those captured by an ordinary cost-benefit analysis. It is emphasised that the findings from such an analysis not be added to the findings from an ordinary cost-benefit analysis, but potentially be included as a supplementary part of the basis for making decisions. Such a supplementary study should, according to the guidelines, be carried out within a relevant transport economic model.

Recommendations are also made with regard to the inclusion in a cost-benefit analysis of any new housing and commercial property development opportunities resulting from infrastructure projects, referred to as “exploitation effects” in Swedish Transport Administration (2012). At the same time, the report notes weaknesses associated with the current approaches to the issue, and it is emphasised that the method must be used with caution to prevent double counting. The findings should under any circumstance not be included in the calculation of net benefits, but may form part of a sensitivity analysis.

7.6 Other sources of discrepancy between realised benefits and estimated economic effects

Arguments relating to wider impacts are often occasioned by a wish not to omit any important effects in the evaluation of projects. The focus of the Committee in the present Chapter is on net wider impacts; effects in other markets than the transportation market with a net impact on economic efficiency as the result of market failure in the secondary market, cf. the discussion of concepts in Chapter 7.2. However, it is also important to discuss whether any aspects of the analysis of impacts in the primary market, which in this case is the transportation market, can be improved. We will therefore in the following examine some sources of discrepancy between realised benefits and estimated economic effects within the transportation market.

Since the benefits from a transportation project largely take the form of reduced travel time, incorrect traffic growth estimates will lead to incorrect realised benefit estimates. Planning takes place under uncertainty, and it cannot be expected that forecasts will always be correct. However, one should expect the forecasts to be unbiased and not, in the long run, to systematically over- or undershoot actual developments. Verification of the priced effects some time after the project has been entered into use, so-called ex-post analysis, may reveal whether the assumptions underpinning the original analysis were correct26. Other types of analysis are required to uncover whether the project triggered net wider impacts, also including other time horizons than would be appropriate for ex-post analysis purposes.

It follows from Kjerkreit and Odeck (2010) that the Ministry of Transport and Communications asked the Norwegian Public Roads Administration, in 2005, to continuously verify the priced effects of projects in progress, and that a guide on the implementation of such verification was issued thereafter. A review of the first eleven analyses conducted within this framework shows that net benefits realised five years after opening exceeded those estimated prior to construction (Kjerkreit and Odeck, 2010). The main reason was that traffic growth had been underestimated at the outset. Other sources of discrepancies between estimates and realised net benefits were over- and underestimation of costs, accident costs and changes in project design subsequent to the original net benefit estimates. Another study, performed by the Institute of Transport Economics, has delivered corresponding findings (Madslien and Hovi, 2007). It follows therefrom that the traffic growth estimates underpinning national transportation plans over the period 1996-2006 were, for the majority of projects, somewhat lower than actual growth. The period examined is too short to merit, in itself, a belief that the forecasting tool is flawed, especially since the global recession in the wake of the financial crisis in 2008 is not included in the data. Nonetheless, the study illustrates the importance of correct forecasts.

In analyses examining what changes have actually been brought about by a project, it is of particular importance to carefully discuss what may actually be considered to be ex ante (before the project has influenced the adaptations of economic agents) and ex post (after the project has influenced the behaviour of concerned economic agents). Many economic adjustments are made immediately upon it becoming realistic to expect the project to be implemented, and well ahead of its physical realisation. If incorrect assessments are made in such regard, one risks not capturing the entire change resulting from the project. One study examines such effects of London being awarded the 2012 Summer Olympics (Brücker and Pappa, 2011). Adaptations in the form of increased investment, consumption and production are identified already prior to the event.

Another source of incorrect evaluation is failure to include factors to which road users or other stakeholders in the analysis attribute value. Let us for example look at a perceived inconvenience associated with dependence on ferry transport to an island. Such inconvenience would then feature as a cost element in the utility functions of economic agents. If an analysis of a bridge or tunnel between the island in question and the mainland ignores such an element, it will have a direct impact on the assessment of costs and benefits, with the result that the benefits such measure will generate for transport users will be underestimated. In addition, it may result in incorrect traffic forecasts, inasmuch as the generalised travel costs faced by agents do not include all costs of relevance to transport users. Bråthen and Hervik (1997) have studied this by looking at five such road links between islands and the mainland. The study found that the realised responses to the projects suggested that the willingness to pay for the new journeys exceeded the original estimates. The reason for this was held to be that transport users were facing an inconvenience cost of ferry transport that was not taken into account in the original analysis, and that one thereby underestimated the real cost reduction the project caused on the part of transport users. This illustrates how ex-post analysis may assist in the identification of omitted cost or benefit elements, such as to facilitate their inclusion in subsequent analyses. The cost element represented by the inconvenience of ferry transport has subsequently been estimated, and is now included in the analysis tool used by the Norwegian Public Roads Administration.

Many effects are, as discussed in Chapter 2, difficult or undesirable to quantify in a cost-benefit analysis. These should nevertheless be included in a systematic analysis of such non-priced effects. If not, relevant non-priced effects will be another source of incorrect evaluation. An example of the systematic analysis of non-priced effects is provided by the analysis of the Sotra link. The project concerns an upgraded road link between Bergen and the island of Sotra, located west of town. The project is accorded priority in the National Transportation Plan 2010-2019, with start-up in the first four-year period and completion in the period 2014-2019. Such priority is conditional upon approval for partial toll funding being granted. The quality assurance (QA1) process established that a concept involving a new two-lane bridge, which, when added to the existing bridge, would result in four lanes, entailed estimated net benefits of NOK -450 million. The quality assurance provider also presented a systematic analysis of the non-priced effects associated with the various alternatives. All in all, the quality assurance provider was of the view that non-priced effects associated with a new two-lane bridge, relating to, inter alia, preparedness for major accidents, as well as improved and more attractive options for pedestrians and cyclists, suggested that the project should be recommended. The reason for the recommendation was that the negative net benefits from monetised effects were deemed to be minor when compared to the overall investment. Moreover, the negative net benefits were held to be compensated for by the positive non-priced implications, especially relating to reduced vulnerability of the road link between the island and the mainland. Another concept involving a new four-lane bridge in the same location, offered net benefits of NOK -1,700 million. Here the non-priced benefits were deemed to be even higher, but not sufficiently high for the quality assurance provider to conclude that these compensated for the lower net benefits from monetised effects. However, the Government has chosen to proceed with the planning of both concepts. The local political choice between the two concepts is now largely focused on balancing the priced against the non-priced effects. The decision on the implementation of the project, and thus on the final choice of concept, has not been made as per completion of the work of the Committee.

Another source of error may be that a project analysis uses an estimate for the average value of time for the country as a whole, despite the journeys undertaken being systematically valued differently from the country average. This is discussed in more detail in Chapter 4.4.3.

In analyses of road pricing in urban areas, it may also be incorrect to use an average value for travel time savings for all road users in a single project. Individual road users are not, generally speaking, faced with the costs they impose on other road users in the form of delays caused by queuing. Road pricing may entail considerable net economic benefits as a result of individual road users being faced with such costs, thus internalising the externalities associated with queuing. However, this reasoning is premised on road users differing in their valuations of time. This means that those who for some reason put a high value on a specific journey at a specific point in time will find it worthwhile to take such journey, even with road pricing. Others will put a lower value on such journey and wait until the road price is lower, for example around mid-day or in the evenings. Anderstig et al. (2011) refer to Parry and Bento (2001), which have examined road pricing from the perspective of the theory on net wider impacts. According to Parry and Bento (2001), net wider impacts in the form of reduced labour supply will be negative when travel costs increase, as with road pricing. However, Anderstig et al. (2011) note that exactly the variation in the valuation of travel time between different road users and purposes of travel suggests that the direction of such correlation is not obvious from the theory. They estimate the effects of the road pricing scheme in Stockholm, and find that net wider impacts of road pricing there were positive.

7.7 The assessment of the Committee

The Committee emphasises that the cost-benefit analysis guidelines must be based on updated research and theoretical developments. It is, at the same time, important for the recommendations to be robust and based on a well-established empirical foundation. This ensures comparability over time and between projects, and confers legitimacy on the analyses.

The approach presented in the NOU 1997: 27 Green Paper with regard to wider impacts is still deemed to be valid. However, new theoretical and empirical research merits some clarifications.

Wider impacts of net economic value

The Committee has focused its discussion on wider impacts that are of neteconomic value to the country, defined as “net wider impacts”. In order for a wider impact to be of net economic value, there must be a market failure in the secondary markets that causes, in the situation prior to implementation of the measure, under- or overconsumption of resources relative to the economically optimal resource allocation. If such under- or overconsumption is affected by the measure under analysis, the said measure has a wider impact that may have an effect on overall economic efficiency.

If goods and services affected by the measure are correctly priced and there is no market failure in the secondary markets prior to the implementation of the measure, effects in other markets will only represent a redistribution of the original effects of such measure, and a well-specified ordinary cost-benefit analysis will capture all economic effects. Such a situation will, in other words, involve no net wider impact. If the redistribution effect is considerable it may nevertheless be of relevance to the assessment of decision makers, and should therefore be addressed in the analysis if possible, cf. the discussion of distribution effects in Chapter 3.

Productivity and economies of scale

It is well-established in economic theory that mechanisms exist that may in certain cases give rise to a positive correlation between productivity and functional town size. This suggests that if a transportation project in an urban area does increase town size in practice, it may result in enhanced productivity for everyone in town, i.e. a net wider impact.

In order to merit a recommendation for the inclusion of effects resulting from such a correlation in cost-benefit analysis, the Committee takes the view that a theoretical and empirical basis should exist for a simple rule that does not depend on a large degree of discretionary assessment in each analysis. This is the only way of ensuring that the analyses are comparable, irrespective of who performs the studies.

However, the review of empirical studies in the present Chapter shows that it is very difficult to identify such a general correlation between functional town size and productivity. Besides, no such studies exist for Norway. The figures reported on how many percent productivity will increase if functional town size increases by one percent vary markedly between sectors, towns and countries (cf., inter alia, Vickerman (2008) and Melo et al. (2009)). The figures also vary markedly depending on which methodological approach is adopted. In addition, a comprehensive overview published by the OECD and the International Transport Forum (Graham and van Dender, 2010) notes that the correlation is not stable and that no correlation at all can be identified for large segments. It is not possible to conclude whether the lack of empirical identification is caused by the theory not working in practice, or whether it means that one has failed to find the appropriate method for distinguishing the effect of town size from other effects that influence productivity.

Since it has turned out to be very difficult to identify a correlation between town size and productivity based on the available data, the Committee takes the view that it cannot generally be recommended to assume such a correlation when preparing estimates for future effects of a project in a cost-benefit analysis context. Part of the difficulty with such identification is that it may take a very long time for the effects to be realised. A cost-benefit analysis should nevertheless discuss, where relevant, whether the type of wider impacts we see discussed in economic theory may arise.

Since economic theory gives reason to believe that there may be positive wider impacts of transportation projects in urban areas, the Committee takes the view that the cost-benefit analysis of large projects related to an urban area where it can be shown to be probable that productivity is systematically higher than in the neighbouring areas may be expanded by a separate discussion of net wider impacts resulting from increased functional town size. It must follow clearly from such an analysis which assumptions the assessments are premised on, and the uncertainty associated with any net wider impacts must be highlighted by way of sensitivity analyses, illustrating the effects of changing the assumptions. The Committee is of the view that any quantitative findings from such a supplementary analysis should not be included in the calculation of the ordinary net economic benefits from a project. This is, firstly, because such effects are highly uncertain and may depend on measures that do not constitute part of the project under analysis. Secondly, including such effects would reduce the comparability of different transportation studies, because the findings from an analysis of net wider impacts will be critically dependent on discretionary assumptions that need to be made in respect of each analysis.

The theoretical arguments in favour of net wider impacts through enhanced productivity resulting from increased functional town size apply to measures in urban areas. This suggests that the theory predicts that any net wider impacts from the construction of a stretch of road in rural areas would be minimal. The empirical review in the present Chapter indicates that the effect of proximity declines steeply with distance, and that no effect can be demonstrated for distances in excess of 50 kilometres from a town centre. When performing a cost-benefit analysis of such projects there is consequently no need to believe that there are net wider impacts as the result of increased functional town size. As far as road links between islands and the mainland are concerned, the situation is so different from place to place that no general recommendation can be made. As with general state aid for industry, it will be difficult to identify which projects may trigger positive net wider impacts. The Committee is not aware of any studies that provide any empirical basis for concluding that integration of housing and labour markets outside urban areas results in any positive net wider impacts for society as a whole.

Labour supply

A tax on labour income means that the economic value of one additional hour worked is higher than the wage referred to by individuals in determining their labour supply. If a project affects overall labour supply in Norway, the additional tax revenues will correspond to the economic value of the net wider impact resulting from distorting taxes. However, it is difficult to know whether individuals will convert reduced travel time into increased labour supply or make other adjustments over time, like for example taking the opportunity to acquire a cheaper home further away from work or to choose a workplace that is further away from their current home. If the benefit is converted into anything other than increased labour supply (and thus increased production), there is no net wider impact as the result of distorting taxes. Moreover, it is not necessarily the case that transportation costs have in practice contributed to individuals not participating in the labour market, and thus that reduced transportation costs will result in increased employment. These are empirical issues. Neither foreign, nor Norwegian, empirical studies exhibit any systematic correlation between overall long-term employment and infrastructure investments. This indicates that one will not be making a major error in excluding such labour supply effects when assessing the economic profitability of transportation projects. As with productivity effects, the identification of such effects may be difficult because it could take a very long time for said effects to be realised in full. A cost-benefit analysis should nevertheless discuss, where relevant, whether the type of wider impacts we see discussed in economic theory may arise.

For major projects where it can be shown to be probable, on an empirical basis, that the project will affect overall labour supply in the country through longer working hours or increased labour force participation, a cost-benefit analysis may, however, be expanded by a separate discussion of these effects. It is in such case important to avoid double counting of the benefits from the project, and the appropriate approach will therefore be to only take into account the tax effect resulting from the change in the value added caused by the increase in employment. However, in order to ensure comparability and in view of the uncertainty, any quantitative findings from such a supplementary analysis should not be included in the calculation of the net economic benefits from a project. The effect of the tax wedge must also be taken into account when valuing the effects of enhanced productivity as the result of increased functional town size.

Land use and transportation

As a general rule, price changes in the property market as the result of a transportation project represent only a redistribution of the original direct benefits from such measure. Consequently, including both effects in the analysis would amount to double counting. If one envisages that a transportation project will enhance productivity in an urban area through the mechanisms discussed in Chapter 7.3.1, one have a situation in which the value of enhanced productivity is reflected in higher property prices. However, if one has sought to calculate the value of enhanced productivity directly, it will also in this case amount to double counting if the effect of the property market price changes is included.

If a measure releases areas that were previously used for transportation, typically by traffic being redirected below ground level, and such areas have a positive value, their value in the best alternative use should in principle be included on the benefit side of the project, inclusive of any option values. However, in order for such an effect to merit inclusion in the analysis there must be a high probability that the areas will actually be used, e.g. in the form of clear expressions of interest from serious stakeholders. Sweden has developed methods for such valuation, but it is recommended that these be used with caution. In any event, the Swedish method recommends that such valuation take the form of a sensitivity analysis, and not be included in the calculation of expected net benefits from the project. The Committee assumes that the Norwegian transportation bodies are monitoring this process closely in their work on cost-benefit analysis guides within their respective areas.

Imperfect competition

Imperfect competition may also be a source of net wider impacts. However, the Committee is of the view that changes in transportation facilities in Norway will in most cases not have any material impact on the degree of imperfect competition. Hence, this should be a limited source of net wider impacts of transportation projects in Norway at present.

In Chapter 7.3.5, it was noted that reduced transportation costs will, when taken in isolation, result in increased production in markets with imperfect competition. This may be a separate source of net wider impacts, also if the competition situation itself is not affected. The magnitude of such an effect will depend on the gap between prices and marginal costs in each sector, and on how the consumers of goods from the affected sector react to the price changes. Individual enterprises will often not have complete information about market conditions and the demand curve in their sector, which suggests a high degree of uncertainty as to whether the market power will be exploited in full and how long time such adjustment will take. Moreover, this will vary from sector to sector. In Norway, there is reason to believe that the composition of commercial transportation will vary considerably between different types of transportation projects, which is an argument against using one single factor to capture such an effect. To the extent that gaps between prices and marginal cost are of some significance, and a firm is facing a downward-sloping demand curve, it is in any event not necessarily the case that reduced transportation cost will in its entirety accrue to the owner, as opposed to employees or subcontractors. This would in such case result in a smaller reduction in the marginal cost of the enterprise than would be indicated by the travel time savings in themselves. On the other hand, if the employees and subcontractors are Norwegians, the portion of the cost savings accruing to them will represent part of the national gain from transportation cost savings. Generally speaking, such a benefit incidence analysis can only be carried out within the framework of a general equilibrium model.

This implies, all in all, that the Committee is not in a position to conclude whether the effect of imperfect competition is of any material importance to the net economic benefits from transportation projects. Besides, it demonstrates that it is difficult to establish a simple method for capturing such an effect in a robust manner based on a firm empirical foundation.

Miscellaneous

It is an important principle that market imperfections should be corrected by policy measures that are as well targeted as possible in addressing the primary cause of such imperfections, such as not to promote enhanced efficiency in one area through measures that give rise to inefficiency in another area. This pertains to imperfect competition, in particular, where the authorities have a more direct policy measure available in the form of addressing market power through the competition legislation and the enforcement thereof. Using transportation investments to increase competition would at best be a second-best solution.27 Valuation in a sensitivity analysis of any net wider impacts resulting from imperfections may in such case illustrate that there is a potential for more direct policy measures to address the imperfection.

Besides, the Committee takes the view that it is no less important to ensure that the specification of the cost and benefit elements is complete, and that there are no other sources of incorrect estimates within the traditional cost-benefit analysis framework. The various sectoral guides must ensure that the main factors in the utility functions of economic agents are included in the analysis and that the valuation takes place at an aggregation level that does not give reason to expected systematically biased outcomes. In choosing an aggregation level, the cost of additional specification must be balanced against the benefits one may derive therefrom. Non-priced effects of importance need to be discussed in a systematic manner.

In practice, cost-benefit analyses are often partial analyses of small projects. Whether a project is small depends on whether it will have a material impact on market prices, cf. the discussion in Chapter 2. If a reform or a measure has a material impact on market prices outside the transportation market, an analysis within a general equilibrium model featuring a realistic transportation network, and taking relevant market failures into account, will be more suitable. Such an analysis framework will not involve the same risk of double counting some effects and omitting other effects as one may run when conducting ad hoc analyses of wider impacts in secondary markets. However, general equilibrium models often operate at a fairly high level of aggregation, which makes them less suitable for capturing the effects of minor changes, for example in the transportation network. The Committee is of the view that the development of general equilibrium models that are better suited for the analysis of such minor changes in the transportation network, and which take the relevant market failure into account, may represent a useful tool for analysing the effects referred to as net wider impacts in the present Chapter.

7.8 Summary recommendations

Based on the discussion in the present Chapter, the Committee makes the following recommendations:

Productivity and economies of scale

  • It has proven very difficult to identify a relationship between town size and productivity when evaluating the effect of a transportation project or a series of such projects subsequent to its or their implementation. According to the view of the Committee it cannot, therefore, be recommended to generally assume such a relationship when evaluating a project prior to its implementation.

  • Since the literature gives reason to believe that there may be positive net wider impacts of transportation projects in urban areas, the cost-benefit analysis of large projects in connection with an urban area in which it can be shown to be probable that productivity is systematically higher, may be expanded to include a separate discussion of net wider impacts. Such an analysis may be both qualitative and quantitative, and should discuss whether such effects are likely to materialise. However, in order to ensure comparability across projects, and in view of the uncertainty, any quantitative findings from such a supplementary analysis should only constitute a supplement to a main analysis of the net economic benefit associated with a project.

Labour supply

  • For major projects concerning which it can be shown to be probable, on an empirical basis, that the project will influence overall labour supply in the country through longer working hours, or through increased labour force participation, a cost-benefit analysis could be expanded with a separate discussion of these effects. Such an analysis may be both qualitative and quantitative, and should discuss whether such effects are likely to materialise. It is important to avoid double counting of the benefits from the project in such contexts, and the correct approach will in practice be to only take into account the change in tax revenues as the result of higher employment. However, in order to ensure comparability, and in view of the uncertainty, any quantitative findings from such a supplementary analysis should only constitute a supplement to the main analysis of the net economic benefits associated with the project.

Land use and transportation

  • As a main rule, price changes in the property market as the result of a transportation project only represent a redistribution of the original direct benefits from such project. Including both effects in the analysis will therefore amount to double counting. If one has sought to estimate the value of increased productivity as the result of increased functional city size directly, it will also amount to double counting if the property market implications of such effects are taken into account. In those cases where a transportation project releases areas with a positive opportunity value, there may be a real economic effect that is not reflected in the direct user benefits of the project.

Imperfect competition

  • Based on the available documentation, the Committee is not in a position to conclude as to whether the effect on imperfect competition is of any material importance to the net economic benefit of transportation projects. The review also shows that it is difficult to establish any simple method for identifying any such potential effect in a robust manner, and with a solid empirical basis. If it can be shown to be probable, on an empirical basis, that the project may influence the degree of competition, or that it will influence markets that are in particular characterised by imperfect competition, a cost-benefit analysis may be expanded to include a separate discussion of these effects. However, in order to ensure comparability, and in view of the uncertainty, any quantitative findings from such a supplementary analysis should only constitute a supplement to a main analysis of the net economic benefits associated with a project.

The ex post analysis of the primary markets

  • The effort to analyse projects subsequent to their implementation should be continued. A systematic approach, like that adopted by the Norwegian Public Roads Administration, generates new knowledge about the analyses carried out, and makes it possible to use these findings to improve the estimates. The systematic follow-up of such studies and other approaches may contribute to ensuring that the specification of the cost and benefit elements is complete, that projections are correct in the long run, and that there are no other sources of incorrect estimates within the cost-benefit analysis framework.

7.9 Bibliography

Anderstig, C., S. Berglund, J. Eliasson, M. Andersson and R. Pyddoke (2011). Congestion charges and the labour market: “wider economic benefits” or “losses”?, Draft CTS Working paper 2011:X. Centre for Transport Studies, Stockholm.

Aschauer, D.A. (1989). Is public expenditure productive? Journal of Monetary Economics, 23, pp. 177-200.

Brücker, M., and E. Pappa (2011). For an Olive Wreath? Olympic Games and Anticipation Effects in Macroeconomics. CEPR Discussion Paper No. 8516.

Bråthen, S. (2000). Do fixed links affect local industry? A Norwegian case study. Journal of transport geography, 9(1), 25-38.

Bråthen, S., and A. Hervik (1997). Strait crossings and economic development. Developing economic impact assessment by means of ex post analyses. Transport Policy, 4 (4), pp. 193-200.

Chatman, D. G., and R. B. Noland (2011). Do Public Transport Improvements Increase Agglomeration Economies? A Review of Literature and an Agenda for Research. Transport Reviews. Vol. 31 (6), pp. 725-742.

COWI (2012). Wider economic benefits of transportation investments. (In Norwegian only. Norwegian title: Mernytte av samferdselsinvesteringer.) COWI report.

Crafts, N. (2009). Transport infrastructure investment: implications for growth and productivity. Oxford Review of Economic Policy, Vol. 25 (3), pp. 327-343.

Department for Transport (United Kingdom) (2005). Transport, Wider Economic Benefits, and Impacts on GDP. Discussion Paper, July 2005.

Department for Transport (United Kingdom) (2012a). Wider Impacts and Regeneration, TAG Unit 2.8, Draft for Consultation, August 2012.

Department for Transport (United Kingdom) (2012b). The Wider Impacts Sub-Objective, TAG Unit 3.5.14, Draft for Consultation, August 2012.

Duranton, G., and D. Puga (2004). Micro-Foundations of Urban Agglomeration Economies in Henderson, J. V., and J.-F. Thisse (eds.), Handbook of Urban and Regional Economics, Volume 4, Cities and Geography, North Holland.

Égert, B., T. Kozluk and D. Sutherland (2009). Infrastructure Investment: Links to Growth and the Role of Public Policies. OECD Economics Department Working Papers, No. 686, OECD publishing, © OECD.

Eriksen, K. S., and V. Jean-Hansen (2008). Better roads in rural areas – do they result in higher business sector productivity? (In Norwegian only. Norwegian title: Bedre veier i distriktene – fører det til høyere produktivitet i næringslivet?) Oslo, Institute of Transport Economics. Working paper ØL/2124/2008.

Fridstrøm, L., and R. Elvik (1997). The barely revealed preference behind road investment priorities. Public Choice 92: pp. 145 –168.

Fæhn, T., M. U. Gulbrandsen and A. Lindegaard (2010). What will be the cost of Norway’s climate cure? (In Norwegian only. Norwegian title: Hva vil Norges klimakur koste?) Samfunnsøkonomen No. 5, 2010.

Graham, D., S. Gibbons and R. Martin (2009). Transport investment and the distance decay factor of agglomeration benefits. Working paper. Imperial College of London.

Graham, D., and K. van Dender (2010). Estimating the agglomeration benefits of transport investments: some tests for stability. Discussion Paper No. 32-2009, OECD/ITF Joint Transport Research Centre.

Hagen, K. P. (2005). Economic Policy and Economic Profitability. (In Norwegian only. Norwegian title: Økonomisk politikk og samfunnsøkonomisk lønnsomhet.) Cappelen Academic Publishers.

Hansen, W. (2011). Wider economic benefits. Wider industrial impacts of infrastructure investments. (In Norwegian only. Norwegian title: Mernytte. Næringsøkonomiske ringvirkninger av infrastrukturinvesteringer.) Institute of Transport Economics (“TØI”) report 1180/2011.

HEATCO (2006). Deliverable 5. Proposal for Harmonised Guidelines. Report, Developing Harmonised European Approaches for Transport Costing and Project Assessment.

Heide, K. M., E. Holmøy, L. Lerskau and I. F. Solli (2004). Macroeconomic Properties of the Norwegian Applied General Equilibrium Model MSG6. Report 2004/18 Statistics Norway.

Heldal, N., I. Rasmussen, S. Strøm and S. Munawar (2009). Wider economic benefits of transportation investments in cities. Planning phase. (In Norwegian only. Norwegian title: Mernytte av transportinvesteringer i storbyer. Forprosjekt.) Vista report 2009/04.

Heum, P., E. B. Norman, V. D. Norman and L. Orvedal (2011). Dryshod to work. Labour market effects of eliminating ferries on the Bergen-Stavanger route. (In Norwegian only. Norwegian title: Tørrskodd på jobb. Arbeidsmarkedsvirkninger av ferjefritt samband Bergen-Stavanger.) Summary. Institute for Research in Economics and Business Administration (“SNF”) 2011.

HM Treasury (2003). Appraisal and Evaluation in Central Government (The Green Book), HMSO, London.

Jara-Diaz, S. (1986). On the Relation between User Benefits and the Economic Effects of Transportation Activities. Journal of Regional Science, 26 (2), pp. 379-391.

Jiwattanakulpaisarn, P., R. B. Noland, D. J. Graham and J. W. Polak (2009). Highway infrastructure investment and county employment growth: a dynamic panel regression analysis. Journal of regional science, 49 (2) 2009, pp. 263-286.

Kjerkreit, A., and J. Odeck (2010). The accuracy of ex-ante benefit cost analysis – a post opening evaluation in the case of Norwegian road projects. Article presented at International Economics Conference (ITrEC) 15-16 June 2009.

Klette, T. J. (1994). Estimating Price-Cost Margins and Scale Economies from a Panel of Microdata. Statistics Norway Discussion Paper No. 130.

Lian, J. I., and J. Rønnevik (2010). Wider impacts of major road projects in Norway. (In Norwegian only. Norwegian title: Ringvirkninger av store vegprosjekter i Norge.) Institute of Transport Economics (“TØI”) report 1065/2010, Institute of Transport Economics.

Madslien, A., I. B. Hovi (2007). Freight and passenger transportation services 1996-2006. Comparison of forecasts and forecast assumptions with actual developments. (In Norwegian only. Norwegian title: Gods- og persontransporttjenester 1996-2006. Sammenligning av prognose og prognoseforutsetninger med faktisk utvikling.) Institute of Transport Economics (“TØI”) report 922/2007.

Marshall, A. (1920). Principles of Economics.

Melo, P. C., D. J. Graham and R. B. Noland (2009). A meta analysis of estimates of urban agglomeration economies. Regional Science and Urban Economics, (39), pp. 332-342.

Minken, H. (2011). Comments relating to wider economic benefits. (In Norwegian only. Norwegian title: Merknader om mernytte.) Working paper ØS/2333/2011. Institute of Transport Economics.

Minken, H., and H. Samstad (2005). Cost-benefit analysis in the transportation sector: Calculation framework. (In Norwegian only. Norwegian title: Nyttekostnadsanalyser i samferdselssektoren: Rammeverk for beregningene.) Institute of Transport Economics (“TØI”) report 798/2005.

Norwegian Public Roads Administration (2006). Impact Assessments. (In Norwegian only. Norwegian title: Konsekvensanalyser.) Handbook 140.

Odeck J. (1996). Ranking of regional road investment in Norway: Does socioeconomic analysis matter? Transportation, (23), pp. 123-140.

Parry, I. W. H., and A. Bento (2001). Revenue Recycling and the Welfare Effects of Road Pricing. Scandinavian Journal of Economics, 103 (4), pp. 645-671.

Proposition No. 50 (2004–2005) to the Storting. Development of the E18 Bjørvika project in Oslo. (In Norwegian only. Norwegian title: Om utbygging av E18 Bjørvika-prosjektet i Oslo.)

SACTRA (Standing Advisory Committee on Trunk Road Assessment) (1999). Transport and the economy: full report.

SIKA (2008). Economic Principles and Estimates for the Transportation Sector: ASEK 4. (In Swedish only. Swedish title: Samhällsekonomiska principer och kalkylvärden för transportsektorn: ASEK 4.) Swedish Institute for Transport and Communications Analysis.

Swedish Transport Administration (2012). Economic Principles and Estimates for the Transportation Sector: ASEK 5. (In Swedish only. Swedish title: Samhällsekonomiska principer och kalkylvärden för transportsektorn: ASEK 5.)

Venables, A. J. (2007). Evaluating Urban Transport Improvements - Cost-Benefit Analysis in the Presence of Agglomeration and Income Taxation. Journal of Transport Economics and Policy, 41 (2), pp. 173-188.

Vickerman, R. (2008). Recent evolution of research into the wider economic benefits of transport infrastructure investments, in OECD/International Transport Forum, The Wider Economic Benefits of Transport: macro-, meso and micro-economic transport planning and investment tools, Round Table 140, Economic Research Centre, OECD, Paris, 2008, pp. 31-49.

Vickerman, R. (2010). Myth and reality in the search for the wider benefits of transport. In Van de Voorde, E., and T. Vanelslander (eds.) Applied Transport Economics: A Management and Policy Perspective. Antwerp: De Boeck, pp. 379-396.

Footnotes

1.

The term “wider economic benefits” is used in many contexts within the transportation sector, but no unambiguous meaning has been attributed to the term. One definition of the term is “benefits that are not included in standard cost-benefit analysis within the transportation sector in Norway at present” (Minken, 2011). This encompasses a very wide range of issues, and will depend on how the analyses are carried out in practice at any given time. Most issues concern direct effects of a measure, which for various reasons are deemed not to be captured to a sufficient extent or with sufficient precision by market prices.. We have addressed some of these sources of error within the transportation sector in Chapter 4.6. The term may also include effects of transportation projects in other markets that have a net economic effect on the country.

2.

This follows from the two main theorems of welfare economics. See e.g. Hagen (2005) for a discussion of these.

3.

See for example Jara-Diaz (1986) for a more detailed discussion of this with regard to the transportation sector.

4.

The presentation in the remainder of this section is largely based on Minken (2011).

5.

The principles behind such gross calculations are presented in more detail in Minken and Samstad (2005).

6.

Specific reference is made to Chapters 17 and 18 of SIKA (2008), Economic Principles and Estimates for the Transportation Sector: ASEK 4. Swedish Institute for Transport and Communications Analysis.

7.

The general equilibrium model, MSG6 (Heide et al., 2004), models market failure as monopolistic competition in the domestic market for a number of goods, as well as direct and indirect taxation that gives rise to a discrepancy between the economic and the private returns on resources. Estimates from this model will therefore take several net wider impacts of a public project/measure into account. MSG6 does not model any detailed transportation network.

8.

At present, Norway has a general equilibrium model that models the geographic distribution of activity; PINGO. However, this model assumes well-functioning markets, and consequently cannot be used to estimate potential net wider impacts (Hansen, 2011).

9.

The study takes the form of an econometric meta analysis of 729 wider impact estimates obtained from 34 different studies.

10.

These figures represent the 5th and 95th percentile, respectively, of the sample.

11.

Crafts (2009) notes that Egert et al. (2009) find significant productivity effects from road and railway investments in the United Kingdom. This is interesting in view of the fact that the UK Department for Transport has promoted the inclusion of measures of such wider impacts in the analysis of major projects in urban areas, cf. guidance on this theme circulated for consultation (UK Department for Transport, 2009a and 2009b). As far as Norway is concerned, no unambiguous statistically significant correlation is identified in Egert et al. (2009).

12.

Graham et al. (2009) define efficient density in a location as.

.

In this equation d is a measure of the distance between location i and location j, whilst zj is employment in location j, and a is the parameter indicating how rapidly effective density declines with distance. If a = 1, effective density is reduced by the inverse of the distance. Graham et al. (2009) have estimated that the mean value a is 1.65, and that it varies between different industries (1.1 for manufacturing, 1.5 for construction, 1.8 for consumer services and 1.7 for producer services). In a consultation paper, UK transportation authorities recommend that these values be used in calculations of wider impacts (Department for Transport, 2012b).

13.

The findings suggest that the correlation between productivity and density is non-linear and that no positive correlation exists between productivity and density for a wide range of density measures.

14.

In their study, the researchers used different econometric methods to examine correlations. They studied freeway densities and employment in 100 counties in North Carolina from 1985-1997, and controlled for other relevant variables that might explain employment. When taking into account the endogeneity of freeway investments and dynamic effects they found no statistically significant correlation between employment and freeway density.

15.

See Department for Transport (2005), paragraphs 191 – 225, for a complete presentation of the reasoning.

16.

Eriksen and Christensen (2001), as reported in Eriksen and Jean-Hansen (2008).

17.

Odeck (1996) and Fridstrøm and Elvik (1997) are studies indicating that transportation investments in Norway are only to a limited extent accorded priority on the basis of economic profitability.

18.

Only a summary of the report is available as per the date of the drafting of the present Green Paper.

19.

Summarised in a lecture Professor Roger Vickerman gave to the Committee on 21 October 2011.

20.

Cf. paragraph 5.25 of HM Treasury (2003).

21.

The effect one intends to measure here is the difference between the willingness to pay for the change in production resulting from the transportation project and the production cost. The recommendation is based on the premise that there is s certain element of market power and that the said difference is positive. Specifically, the recommendation calls for a ten percent increase in the direct user benefits originating from business travel and the conveyance of goods.

22.

Here the UK Department for Transport presents a method for estimating the change in employment resulting from a reduction in generalised travel costs, and thereafter recommends including the tax wedge associated with the change in employment in the analysis, which according to economic theory (cf. for example Venables, 2007) will not be captured by the direct user effects.

23.

Here the UK Department for Transport presents a method for estimating the effect. It calls for estimates to be prepared as to the change in efficient accessibility for a number of small zones in the area that is being studied, as well as for subsequent estimates as to the potential productivity effect thereof, with the economy being divided into four defined industries. The UK Department for Transport has prepared a table setting out recommended values for the elasticity of productivity with respect to efficient density.

24.

Findings by the Department for Transport (2005), as reported by Crafts (2009).

25.

Odgaard, T., Kelly, C.E., and Laird, J.J. (2005), Current practice in project appraisal in Europe – Analysis of country reports, HEATCO Work Package 3. HEATCO – Developing Harmonised European Approaches for Transport Costing and Project Assessment. Funded by the 6th Framework RTD Programme, IER Stuttgart, Germany.

26.

Such verification may also be considered in the context of the evaluation requirement stipulated in Section 16 of the Central Government Financial Management Provisions.

27.

Vickerman (2008) notes that the competition effect of transportation investments can under any circumstance be expected to be neutral. On the one hand, reduced transportation costs may expand the market, thus introducing new competitors. On the other hand, such increased competitive pressure may in practice lead to some firms being bankrupted and thus reduce the number of firms competing in the market in the long run. It is noted that the exception will be those cases where a road link does in itself contribute to “unlocking” a previously protected local monopoly. Vickerman believes that this will rarely be the case in modern economies.

Go to front page